Perlin Noise Creating Procedural Solid T extures Student: - - PowerPoint PPT Presentation

01/07/10 Presentation 1

Perlin Noise

Creating Procedural Solid T extures Student: Alexandre Chapiro, Advisor: Luiz Velho

2 Presentation 01/07/10

T

• pics

Topics Theory Some examples

• Solid textures.
• Perlin Noise.
• What can we do with all this?

Software Introduction More examples!

3 Presentation 01/07/10

Introduction

Introduction Theory Some examples

➙ What is Perlin Noise?

• Perlin Noise is a method used to develop several

interesting procedural textures:

• Clouds / Water / Fire / some materials (marble,

wood) ...

• ... and much more!

Software Topics More examples!

4 Presentation 01/07/10

Theory

Theory Introduction Some examples

➙ What is Perlin Noise?

• Perlin Noise is a mapping from ℝn to ℝ.
• First, create a grid of points with integer

coordinates,

• For each point, find it's closest neighbours on this

grid,

• Use a hash table to assign each of the grid points

a gradient vector, which defines a linear function,

• Weight these functions using an S shaped

polynomial (currently 6x⁵ – 15x⁴ + 10x³),

• Sum them all, and the result is Perlin Noise.

Software Topics More examples!

5 Presentation 01/07/10

Theory

Theory Introduction Some examples

➙ What does it look like?

• Perlin Noise alone isn't

very useful.

• We can, however,

combine it with several

• ther functions!
• For instance, ⅀(1/f)*(noise(f*p))

is called turbulence, and can be used to generate lots of interesting effects.

Software Topics

↑This

More examples!

6 Presentation 01/07/10

Some examples

Some examples Theory Introduction

➙ Some examples from Perlin's “Making Noise” talk (and my attempts at replicating them):

Software Topics

➙Original Image ➙My example

More examples!

7 Presentation 01/07/10

A water ball with ripples

Some examples Theory Introduction

➙ How was it obtained?

Software Topics

• Only bump mapping, no color

variation;

• The normal vector at each point

'p' is obtained through: n += pnoise(p*parameter);

• 'n' is the original normal;
• Parameter is chosen big enough

for a pleasing effect. (more on this later)

More examples!

8 Presentation 01/07/10

Some examples

Some examples Theory Introduction

➙ Another example from Perlin's “Making Noise” talk:

Software Topics

➙Original Image ➙My example

More examples!

9 Presentation 01/07/10

A green-ish planet with clouds

Some examples Theory Introduction

➙ How was it obtained?

Software Topics

• Outside of the ball brightened, to

make it look like a clearer atmosphere arround a darker landmass;

• Planet's color is static, with

clouds added to the original color;

• Cloud's color obtained with:

c += turbulence (8, p);

• 8 is the amount of octaves;

More examples!

10 Presentation 01/07/10

Some examples

Some examples Theory Introduction

➙ Yet another example from Perlin's “Making Noise” talk:

Software Topics

➙Original Image ➙My example

More examples!

11 Presentation 01/07/10

Flame ball

Some examples Theory Introduction

➙ How was it obtained?

Software Topics

• Initial colour set as bright orange

(lots of R and G);

• The color is then obtained by

keeping R the same and varying G:

G -= 4* [turbulence (15, parameter*p, WITH_ABSOLUTE_VALUE)]^2;

• 15 octaves, turbulence with

absolute value, ^2 for thinner lines.

More examples!

12 Presentation 01/07/10

Some examples

Some examples Theory Introduction

➙ One last example from Perlin's “Making Noise” talk:

Software Topics

➙Original Image ➙My example

More examples!

13 Presentation 01/07/10

Marble

Some examples Theory Introduction

➙ How was it obtained?

Software Topics

• Initial colour set as “hay”, then

multiplied by 'f';

• K = turbulence(15, p,

WITH_ABSOLUTE_VALUE);

C = 'hay'* sin(4*PI*(p.x + 2*k^2);

• 15 octaves, turbulence with

absolute value, ^2 for thinner lines, 4*PI sets the ammount of

• scillation of the 'sin' function

(black x colored lines).

More examples!

14 Presentation 01/07/10

The program.

Theory Some examples Introduction

➙I've created an interactive program that

allows the user to create his own textures using Perlin Noise / T urbulence.

Topics More examples! Software

15 Presentation 01/07/10

Normal editing.

Theory Some examples Introduction

➙The first tab allows the user to create bump

maps using diferent kinds of T urbulence, stripes, and several possibilities.

Topics More examples! Software

16 Presentation 01/07/10

T exture creation.

Theory Some examples Introduction

➙Users can create their own textures. Some examples:

Topics More examples! Software

17 Presentation 01/07/10

Defines Regions.

Theory Some examples Introduction

➙The program also allows the user to create a division on the surface using turbulence or stripes...

Topics More examples! Software

18 Presentation 01/07/10

The final result!

Theory Some examples Introduction

➙Finally, you can create an image that contains the respective textures in the created regions,

• n a sphere with the bump map you used

before!

Topics More examples! Software

01/07/10 Presentation 19

More results!

01/07/10 Presentation 20

Few more results!

21 Presentation 01/07/10

References:

• 1) Ken Perlin, “An Image Synthesizer”, SIGGRAPH

Volume 19, number 3, July 1985.

• 2) Ken Perlin, “Improving Noise”, Proceedings of ACM

SIGGRAPH 2002.

• 3) A. Lagae et al., “State of the Art in Procedural

Noise Functions”, EUROGRAPHICS, STAR – State of The Art Report, 2010.

• Ken Perlin, “Making Noise”,

http://www.noisemachine.com/talk1

Papers: Talks:

01/07/10 Presentation 22

Thank You!